Organic farming practices are regarded as being beneficial for the environment by promoting soil quality and sequestering soil organic carbon (SOC). We studied SOC dynamics in the long-term field experiment DOK in Switzerland. The experiment compares three organically fertilized treatments under conventional (CONFYM), bioorganic (BIOORG), and biodynamic (BIO-DYN) management, and two systems with (CONMIN) or without (NOFERT) mineral fertilizer. We analyzed measured SOC time series from 1977 to 2004 and applied soil fractionation, radiocarbon dating, and modeling with the carbon model RothC. The SOC declined significantly in most parcels, but was not systematically different between systems. Initial SOC contents correlated with soil texture and were identified as being important with respect to the change rate. The SOC loss was at the expense of mineral-associated carbon whereas the more labile fractions increased. The overall decline was explained by reduced carbon inputs since commencement of the experiment and was most pronounced in NOFERT and CONMIN. The model satisfactorily simulated the dynamics of most of the treatments for both initialization with equilibrium runs or measured SOC fractions. Carbon loss in CONFYM was not fully captured by the model. Composition of organic fertilizers depended on the particular management, and a model adjustment of their relative stability improved the match between model and measurements. Model runs without management effects indicated that the observed increase in temperatures at the experimental site does not induce a change in SOC. Overall, the study does not support a benefit of organic farming on SOC contents compared with conventional systems with manure.
